Ceylon Migration Bot

[@lucaswerkmeister] Okay, but `north {}` also isn’t the valid Ceylon expression to get `north` – that would be `north`, which is exactly what you want. So, in general, the generated...

[@lucaswerkmeister] > I don't mind, it's just text, but it _does_ seem to be an instantiation which it is not. Well, for me, it’s _supposed_ to be a valid instantiation...

[@quintesse] > Well, for me, it’s supposed to be a valid instantiation Because it's very well possible that the instantiation arguments have nothing to do with the attributes you're going...

[@lucaswerkmeister] > Because it's very well possible that the instantiation arguments have nothing to do with the attributes you're going to mark as `identifying`, so even if it looks like...

[@lucaswerkmeister] > Syntax? There's no syntax here, it's just a text meant for humans to read. Okay, format then, or something. It’s a regular format the compiler will be generating,...

[@gavinking] Look, if there are no attributes marked for inclusion in the `string`, then the format is just the class (or object) name. Easy.

[@quintesse] > then the format is just the class (or object) name Which is what I said ;) > perhaps the annotation shouldn’t be on the individual attributes, but on...

[@lucaswerkmeister] > We could have both. A class-level one to indicate we want those things to be generated and then optional field-level ones to indicate which ones are to be...

[@quintesse] > you shouldn’t present string to the end user, right? in general, right, more for debugging and logging, although our built-in basic value types like Integer and String (and...

[@lucaswerkmeister] @renatoathaydes [suggested](https://groups.google.com/d/msg/ceylon-users/R2q5LpOD7Eo/Cf1QuAQeK2EJ) that this could be added using macros (#1865), which I think is a really elegant way to do it (no extra complexity in the compiler itself).